In Sindarin and Noldorin, any final [v] after the high vowel [i] became [w], as noted by David Salo (GS/§4.173, GS/§4.174), though he posited that this sound change occurred earlier, before [ṽ] and [β] (from [m] and [b]) blended into [v]. The change of [v] > [w] also occurred after the diphthongs [ae] < [ai] and [oe] < [oi], and it is very tempting to assume (as David Salo did) that the change of final [v] > [w] occurred before [ai], [oi] became [ae], [oe]. Unfortunately, we have a Noldorin counter-example (published after Salo’s Gateway to Sindarin) that clearly shows [ai] > [ae] before [v] > [w]:

• ON. sagma > saʒmh > sae̯mh > saęw “poison” (PE22/32).

Aside from some ambiguity on the ordering of the sound changes, there are plenty examples to make it clear this sound change occurred for [v] of various origins after [i], [ae], [oe]:

• ✶wagmē [> *(g)waiv] > S. gwaew “storm” (NM/237).
• ✶okma [> *oiv] > S. oew “evil deed” (PE17/170).
• ᴹ✶laibē > ON. glaibe [> *glaiv] > N. glaew “salve, ointment” (PE22/32; Ety/LIB²).
• ᴹ✶khīmā [> *hīv] > N. hiw “sticky, viscous” (Ety/KHIM).

The word S. raef “net” < *raima from the late 1960s is a possible counterexample (PE22/159), but this might just be an archaic form since Tolkien gave both raef, raew “net” in another note from the late 1960s (VT42/12).

Conceptual Developments: In Gnomish of the 1910s the sound change [m] > [v] did not occur, so there were far fewer potential examples of [v] > [w], making it difficult to analyze this sound change. There is even one example that seems to hint at a reversed development of [w] > [v] or [f]: ᴱ✶duiwe > G. duif “stream”, extended forms duivrad “canal” and duivrant “aqueduct” (GL/31). Lack of examples make the phonetic developments of the Early Noldorin of the 1920s hard to analyze as well.

GS/§4.173 GS/§4.174 but S. raef @@@